The subject matter relates to the products and semi-finished products of the tobacco industry. Products of the tobacco industry, such as cigarettes and filter rods as well as the semi-finished products such as sections of cigarettes and filter rods are defined as rod-like articles or rod-like products. They are manufactured or processed by means of the line techniques or drum techniques. In both techniques finished products are made as a result of cutting an endless rod formed by means of a wrapper, whereas the endless rod consists of various types of tobacco and filtering materials, and some other materials. Such a rod is cut into single articles which are fed to successive stages of the manufacturing process. As regards the line technique, the component parts are put together with each other on a moving band. However, in the case of the drum technique, the component parts, for instance filter rods, are put together with each other on drum conveyors, and subsequently wrapped in a wrapper. Regardless the applied technique, the rod-like articles undergo an inspection, during which the quality parameters of the manufactured rod-like articles are checked, in order to reject defective articles from the manufacturing process.
According to state of the art, there are known documents presenting various systems for performing quality inspection of the manufactured rod-like articles. The patent U.S. Pat. No. 4,212,541A discloses the manufacturing system for producing multi-segment rods equipped with a measuring unit and a system for rejecting defective products.
Manufacturers check many different parameters which determine the articles quality, whereas for various articles, various parameters can be checked. For the filter rod manufactured of single type of filtering material, the quality parameters are diameter, length, materials uniformity, and lack of foreign inclusions. In case of the filter rod manufactured of the uniform material, in which the capsules are inserted (for example with the aromatic substance), additionally the following parameters will be checked: presence of capsules in proper places, capsule size, capsule quality, distance of a capsule from the rod end, distances between particular capsules as well as the central position of the capsule in the lateral cross section of the rod. In case of the multi-segment rods consisting of various segments, the crucial quality parameters will be the following: lengths of particular segments, distances between particular segments, and if the segments consist of capsules, then, additionally, position of capsules in segments along the axis as well as in particular cross sections of segments. During the manufacturing of the rod-like articles, at least few quality parameters are checked, whereas the measurements of the parameters are conducted at various speeds, within the entire efficiency range of the manufacturing machines. There is a number of crucial aspects of conducting measurements. The most important aspect is performing measurements at high speed of the manufacturing machine. For the speed of an endless rod of the order of 500 m/min the risk of conducting inaccurate measurements increases. The dynamics of the process increases so does the risk of measurements interferences. The measurement time is very short and the risk of effect of random factors, which are not taken into account by the control system and lead to measurement errors, increases. In order to ensure high efficiency of the measuring unit, measurements of certain parameters are performed by means of several sensors and the measurements results from each sensor are compared to each other or averaged. Alternatively, one parameter can be measured by means of multiple sensors measuring the same parameter on the basis of various principles of measurement, and then the results are compared to each other. The quality parameters have predetermined ranges, within which the values of the measured parameter should fall. The control system receives the results of all measurements and takes the decision to reject an article if the actual value of any of the parameters does not fall within the required range of the acceptable parameter values. There is a demand of the tobacco industry manufacturers to reject only such articles, the quality parameters of which indeed fall outside the acceptable limits of parameters, i.e. outside the parameter tolerance limit. The most serious problem during conducting measurements is the fact that part of the parameters is close to the limit values of the parameter range. The measurement of the value of the parameter by the measuring unit has an accuracy of a measuring sensor. Therefore, close proximity of the measured value to the acceptable limit of the parameter increases risk of incorrect article classification. Low accuracy of the measurement introduces the risk that the measurement of both good and defective article can produce the same result. In the case of conducting the measurement of one parameter by means of two or three sensors and if there are results discrepancies, i.e. when some of sensors indicate that the measured article fulfils the quality requirements, whereas the remaining sensors indicate that the measured article does not fulfil the quality requirements, there is a risk of incorrect article classification. Aspiring to eliminate all potentially defective articles from manufacturing, all the articles, the parameters of which cannot be clearly defined, can be rejected. Nevertheless, there is a demand among the manufacturers for a system which will be rejecting from manufacturing only defective articles and which will not be rejecting articles having the correct quality parameters.